Understanding Child Maltreatment Report Risks: A 14-Year Longitudinal Study

Background/Purposes:

The study of child maltreatment, as evidenced in official child maltreatment reports (CMR), has burgeoned in the past two decades. There are few studies, however, examining CMR longitudinally. This study fills that gap by examining CMR at each child age during a 14-year follow-up while simultaneously considering changes of various risk/protective factors over-time.

Methods:

We followed up two separate samples from 1991-1994 St. Louis birth cohorts representing two different high-risk populations over a 14-year period. The “CAN” sample included all infants/toddlers with a first-time CMR in 1993-1994 (n = 2,111). The “AFDC” sample included randomly selected infants/toddlers receiving AFDC in 1993-1994 with no current/prior CMR (N = 1,923). We followed them through 2009 and estimated the likelihood of having a CMR at each age from 1-17. We used multilevel logistic growth curve models to estimate CMR likelihood as a function of various predictors. We measured variables by tracing children in various Missouri administrative records.

Results:

There was a CMR for 11.5% of child ages (CAN sample) and 5.5% of child ages (AFDC sample). In other words, the CAN sample had an about an 11.5% chance of being reported during each year we observed them. Multivariate models revealed important risk/protective factors predicting CMR. Every 1-year increase in child age decreased CMR likelihood by 13% for the CAN sample (Odds Ratio=0.87, 95% CI=0.85-0.88) while child age was non-significant for AFDC sample (0.98, 0.95-1.01). Current welfare receipt (receiving AFDC/TANF at current age) increased CMR likelihood by 98% for CAN-sample (1.98, 1.73-2.26) and 67% for AFDC-sample (1.67, 1.40-2.00). Welfare receipt history (% of months on AFDC/TANF from birth to prior age-year; 1 unit = 10-percentage point) increased CMR likelihood by 8% for CAN-sample (1.08, 1.05-1.11) and 12% for AFDC-sample (1.12, 1.08-1.17). The effect of current welfare receipt significantly decreased with an increase of welfare receipt history for both samples and with an increase of child age for AFDC sample. In both samples, Blacks had significantly “lower” CMR likelihoods than Whites after controlling for other predictors, especially welfare receipt. Many other predictors including prior CMR, child mental health, child injury, child special education needs, parent criminal behaviors, parent low education, and mother’s own foster care placement increased future CMR in both samples. Some predictors were significant only for the CAN sample: receiving preventive (in-home) services lowered future CMR and child behavioral problems increased future CMR. Baseline neighborhood impoverishment and instability were non-significant in both samples.

Conclusions/Implications:

Results suggest that low-SES remained a strong predictor of future CMR risks even among high-risk children. The strong observed interactions of low-SES with child age and cumulative exposure to low-SES suggest the importance of longitudinal approaches in understanding their relationships to CMR. Implications include the centrality of poverty in CMR, and necessity of addressing this critical issue in policy and practice. Additionally, this study highlights the utility of cross-sector data in improving our ability to better understand and predict child maltreatment.